prompt-maker-studio / backend
AGENTS.md7 starsChanged 18 days agoApache-2.0View on GitHub
FastAPI. Layering is api/ → services/ → models/ → database/; route handlers stay thin and delegate to services/. Boundary schemas live in domain modules under schemas/. REST and MCP share services/prompts/; map OperationError through the transport adapter instead of calling a route handler from another transport. Use the add-api-endpoint skill for new routes. LLM calls go through services/llm_client.py and nowhere else. It is the single place an OpenAI(...) client is constructed, resolving the acting user's own bring-your-own provider connection, or…
- Reads credentials
Where files are headingNew AGENTS.md files have more rules about keeping secrets out of the code, and fewer about code style.
Of the AGENTS.md files we read, the 31,439 first seen in the last six months, against the older ones: the kind of rule that gained most share, and the kind that lost most. About files like this in general, not this one, and an estimate, since a kind is a pattern in the words.
In this repository10 other files it gives its agents.
- AGENTS.mdAGENTS.md
- frontend/AGENTS.mdAGENTS.md
- add-api-endpointSkill
- code-reviewSkill
- coding-agent-briefSkill
- db-migrationSkill
- find-skillsSkill
- impeccableSkill
- product-ownerSkill
- readme-syncSkill
The file
We split this AGENTS.md into the rules it gives an agent
1 rule. About 0 are common, 0 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 15
Don't hand-format — run make lint-backend, or make lint-fix to auto-fix (ask before running lint-fix, since it mutates source files).
≈ “Common” is an estimate: at least 1 in 5 of the 40,246 AGENTS.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
backend/ — agent operating notes
FastAPI. Layering is api/ → services/ → models/ → database/; route handlers stay thin
and delegate to services/. Boundary schemas live in domain modules under schemas/.
REST and MCP share services/prompts/; map OperationError through the transport
adapter instead of calling a route handler from another transport. Use the add-api-endpoint skill for new routes.
LLM calls go through services/llm_client.py and nowhere else. It is the single place an OpenAI(...) client is constructed, resolving the acting user's own bring-your-own provider connection, or the operator's server default (LLM_PROVIDER, LLM_BASE_URL, LLM_API_KEY, LLM_MODEL, LLM_WORKSPACE_ID, LLM_TIMEOUT_SECONDS, read only there) when they have none. LLM_PRICING_* belongs to services/llm_pricing.py and LLM_ENCRYPTION_KEY to services/secret_store.py. AI services take an LLMConnection argument (EvalService.run_evaluation resolves one with client_for(user) for its workers); they never read credentials from the environment. Anything needing JSON back uses json_completion(), which handles the fact that response_format is not portable across OpenAI-compatible endpoints. See product/DECISIONS.md, "Bring-Your-Own LLM Provider over an OpenAI-Compatible Transport".
No Alembic. Schema changes go through the idempotent runner in app/database/migrations/runner.py, not migration-tool commands or raw SQL against a live DB. The numbered .sql files in backend/migrations/ are historical references only, not directly executed. See backend/migrations/README.md for the full mechanism and the production migration checklist. Use the db-migration skill when making a schema change.
Register new models in app/models/__init__.py so standalone scripts resolve the same
ORM relationships as the API.
Don't hand-format — run make lint-backend, or make lint-fix to auto-fix (ask before running lint-fix, since it mutates source files). pyproject.toml is the single source of truth for Ruff, pytest, and coverage settings; there is no pytest.ini or .coveragerc.
Dependencies are declared in backend/pyproject.toml and locked transitively in backend/requirements.lock (runtime, used by the Dockerfile) and backend/requirements-dev.lock (adds dev tooling, used by CI and make setup-venv). After changing any dependency, run make lock and commit both lock files with the change.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

